Extreme categories have one configuration while gray-zone cases have many

Evidence

While one should always be cautious about generalizing from one or two examples, my intuition is that often there are lots of ways to be a competitive–authoritarian regime. More generally, one reason why data sets often differ on coding the gray zone is that there are lots of configurations of attributes in it. To use the Polity concept of democracy–autocracy, at the extremes there is only one way to be a complete democracy or complete autocracy, however there are many combinations that put a state in the middle. — Gary Goertz, Social Science Concepts and Measurement, 1134

Categorization

• XREF: Connects to the Milgrom law-merchant material on why coding rules need to handle many configurations in the middle, and to fuzzy-logic set-membership discussions earlier in the book.
• SEED: Angle: ‘Why experts can’t agree on the same country’s democracy score’ — the gray zone’s combinatorial explosion as the real source of inter-coder disagreement.
• QUESTION: Is it generally true that extremes are singular and middles are multiple, or only for some attribute spaces (e.g. institutional rather than continuous traits)?